More than a dozen SNPs have been found to be associated with prostate cancer (PCa) risk by genome-wide association studies (GWAS). These reported PCa risk associated SNPs, if not driven by PSA detection bias or not completely correlated with PSA, could be used to improve PSA and other existing clinical variables in predicting positive prostate biopsy (i.e. PCa). These two important questions, however, cannot be addressed by most PCa case-control studies because some cases were diagnosed on the basis of elevated PSA levels. Only studies such as the Prostate Cancer Prevention Trial (PCPT) where men at the end of the trial were biopsied regardless of PSA levels can be used to dissect these two questions. The overall hypothesis of the study is that multiple genetic variants, when combined, can be used to predict men at increased risk for PCa. Specifically, we hypothesize that 1) a subset of genetic variants are associated with PCa risk and are not solely due to PSA detection bias, 2) these genetic variants, when combined, are strongly associated with PCa risk, 3) genetic variants can supplement PSA and other existing clinical variables to improve the predictive performance for PCa and aggressive PCa, and 4) the chemoprevention effect of finasteride is different among men with different genetic risks and the reduction in PCa diagnosis by finasteride is larger for men with higher genetic risk. To test these hypotheses, we will use data and samples from the PCPT study, a phase III randomized, double-blind, placebo-controlled trial of finasteride in the prevention of prostate cancer. We have three specific aims. Aim 1 is to test whether reported prostate cancer risk associated variants from GWAS are associated with PCa risk, free of PSA detection bias. Aim 2 is to estimate the joint predictive performance for prostate cancer diagnosis, overall PCa and aggressive PCa, using genetic variants as well as PSA and other existing clinical variables. Aim 3 is to assess the differential chemoprevention effect of finasteride on prostate cancer diagnosis among men with higher or lower genetic risk for PCa. Results from this study may potentially benefit millions men. Men at highest risk for PCa could be identified at an early age for intensive screening and chemoprevention such as finasteride. Genetic variants could also be used in combination with PSA and other existing clinical variables to considerably improve their predictive accuracy for positive prostate biopsy. PUBLIC HEALTH RELEVANCE: Dozens of prostate cancer risk associated variants, when combined, could be used to identify men at highest risk for prostate cancer. Such men could be identified at an early age for intensive screening and chemoprevention. Prostate cancer risk associated variants could also be used in combination with PSA and other existing clinical variables to considerably improve their predictive accuracy for positive prostate biopsy. 1